NK.
All work

Case study 06 of 08Complete2025

Insurance Analytics PoC

Insurance analytics. Semantic models, with an AI assistant on top.

Insurance analytics PoC with synthetic datasets, semantic models, Power Query/DAX transformations, in-report Copilot, prompt libraries, verified response patterns, and a parameterized SQL data agent.

Role
Analytics PoC builder
Synthetic data, semantic model, answer validation
Stack
Power BI + Fabric
DAX, Power Query, SQL patterns, Copilot-assisted analysis
Scale
Synthetic domain model
Insurance-style measures and prompt workflows
Outcome
Checkable answers
Prompt libraries and SQL patterns make responses reviewable
01

The brief. What it took, and what it changed.

Problem

Demonstrate trusted analytics experiences where users can ask natural-language questions and receive verifiable, data-grounded answers.

System

Synthetic data feeding Power BI semantic models and report experiences, paired with prompt libraries, response verification, and parameterized SQL data-agent access.

Impact

Supports the hybrid story across BI, semantic modeling, and data-agent experimentation.

02

How it works. In 3 stages.

    01

    Synthetic dataset

    Sanitized data keeps the example reviewable without exposing client or operational records.

    02

    Semantic model

    Power Query, DAX, and curated measures ground natural-language answers in trusted BI logic.

    03

    Verified responses

    Prompt libraries and parameterized SQL patterns make generated answers easier to check.

03

What shaped the build. And what it runs on.

    Use as supporting work for AI-assisted analytics and semantic-model craft.

Built with

  • Power BI
  • Power Query
  • DAX
  • SQL
  • Copilot
  • data agent patterns
04

Proof. What's public, and what's on its way.

References

  • PoC referenceScreenshots, prompt examples, or response verification samples to add after approval.Pending

Visual artifacts

  • PoC screenshotsscreenshotTo approve
  • Prompt-library sampledocumentTo approve
  • Response verification exampledocumentTo approve
05

Tradeoffs. And what stays private.

The tradeoff

Keep below the top three project cards unless strong visuals or demo artifacts are collected.

Kept anonymous

Confirm whether PoC details and visuals can be public.

Next case studyVITAC. Sizing a skincare market, then telling its story.Read it